Optimize Document Processing Performance

[aybanda]

Resolves #2 - High latency during document processing under moderate to high load

Changes

• Implemented parallel processing with dynamic worker allocation
• Added caching for extracted text to avoid redundant processing
• Optimized resource management and batch processing
• Added comprehensive performance testing suite

Performance Improvements

• Low load (2 users, 10 docs): 89.19% faster
• Medium load (5 users, 20 docs): 87.65% faster
• High load (10 users, 30 docs): 70.69% faster
• Maximum latency reduced from 3.21s to 0.98s under high load

Testing

• Added test_performance.py for comprehensive load testing
• Included benchmark results in benchmark_detailed_results.json
• Validated improvements across different document complexities
• Successfully handles concurrent user scenarios

Files Changed

• Unsiloed/services/chunking.py: Core implementation changes
• test_performance.py: Performance test suite
• benchmark.py: Benchmarking utility
• benchmark_detailed_results.json: Performance metrics

/claim #2